Getting Started
Start with the quickstart: it runs the engine in Docker and an order workflow in Rust, Python or TypeScript, then kills the worker mid-run so you can watch the run resume from its last completed step.
Quickstart
Start the engine, run a worker, start a workflow, then kill the worker and watch the run resume.
Architecture
How the engine, PostgreSQL and your workers fit together, and how a run is journaled and resumed.
Features
What the engine does: durable execution, timers, events, child workflows, actors and the SDKs.
Choose your SDK
Each SDK is pre-1.0: its API may change between minor releases, and its changelog lists what changed.
Rust
#[workflow] and #[task] macros, typed task references, and registration at startup with no manual wiring.
$ cargo add orcher-sdkPython
@workflow and @task decorators on plain async functions, a worker builder and a typed client.
$ pip install orcher-sdkTypeScript
Functional or decorator-based definitions, typed task references, and dependency injection for tasks.
$ npm install @orcher/sdkUse cases
The kinds of systems durable workflows are used for.
Data Pipelines & ETL
Pipelines that resume from the last completed batch, and parallel transforms.
Payment Processing
Payments, refunds and reconciliation that must not repeat a charge.
Microservice Orchestration
Child workflows, event gates, and saga patterns for consistency.
Background Jobs
Email campaigns, scheduled reports, and async workloads with retries.
Event Processing
Event-driven pipelines, stateful actors, and pattern detection.
Long-Running Processes
Approvals, trial management, recurring tasks, and order fulfillment.
Example projects
Clone, run, and learn from working code.